Long COVID and Hybrid Immunity among Children and Adolescents Post-Delta Variant Infection in Thailand.

This study aimed to assess long COVID, and describe immunogenicity against Omicron variants following BNT162b2 vaccination. A prospective cohort study was conducted among children (aged 5-11) and adolescents (aged 12-17) who had SARS-CoV-2 infection from July to December 2021 (Delta predominant period). Long COVID symptoms were assessed by questionnaires at 3 months after infection. Immunogenicity was evaluated by using a surrogate virus-neutralizing antibody test (sVNT) against the Omicron variant. We enrolled 97 children and 57 adolescents. At 3 months, 30 children (31%) and 34 adolescents (60%) reported at least one long COVID symptom, with respiratory symptoms prevailing (25% children and 32% adolescents). The median time from infection to vaccination was 3 months in adolescents and 7 months in children. At 1 month following vaccination, in children who received one-dose and two-dose BNT162b2 vaccines, the median (IQR) sVNT against Omicron was 86.2% inhibition (71.1-91.8) and 79.2% inhibition (61.5-88.9), respectively (p = 0.26). Among adolescents who received one-dose and two-dose BNT162b2 vaccines, the median (IQR) sVNT against Omicron was 64.4% inhibition (46.8-88.8) and 68.8% inhibition (65.0-91.2) (p = 0.64). Adolescents had a higher prevalence of long COVID than children. Immunogenicity against the Omicron variant after vaccination was high and did not vary between one or two doses of the vaccine in either children or adolescents.

Neurological complications associated with influenza in hospitalized children.

BACKGROUND: Influenza is a known respiratory and potential neurotropic virus. This study aimed to determine the prevalence and outcomes of influenza-related neurological complications among hospitalized children. METHODS: All medical records of hospitalized children aged <18 years old diagnosed with influenza at a tertiary care hospital in Bangkok were retrospectively reviewed. Influenza infection was confirmed by rapid antigen or reverse transcription polymerase chain reaction tests. Neurological characteristics and clinical outcomes were analyzed using the Pediatric Cerebral Performance Category Scale. RESULTS: From 2013 to 2018, 397 hospitalized children with a median age of 3.7 years (interquartile range [IQR]: 1.6-6.9) were included. The prevalence of neurological complications, including seizure or acute encephalopathy, was 16.9% (95% confidence interval [CI]: 13.3-20.9). Influenza A and B were identified in 73.1% and 26.9% of the patients, respectively. Among 39 (58.2%) acute symptomatic seizure cases, 25 (37.3%) children had simple febrile seizures, 7 (10.4%) had repetitive seizures, and 7 (10.4%) had provoked seizures with pre-existing epilepsy. For 28 (41.8%) encephalopathy cases, the clinical courses were benign in 20 (29.9%) cases and severe in 8 (11.9%) cases. Ten (14.9%) children needed intensive care monitoring, and 62 (93.5%) fully recovered to their baselines at hospital discharge. Predisposing factors to the neurological complications included a history of febrile seizure (adjusted odds ratio [aOR]: 20.3; 95% CI: 6.6-63.0), pre-existing epilepsy (aOR: 3.6; 95% CI: 1.3-10.2), and a history of other neurological disorders (aOR: 3.5; 95% CI: 1.2-10.2). CONCLUSIONS: One fifth of hospitalized children with influenza had neurological complications with a favorable outcome. Children with pre-existing neurological conditions were at higher risk for developing neurological complications.

Immune Response after 2 Doses of BNT162b2 mRNA COVID-19 Vaccinations in Children and Adolescents with Cancer and Hematologic Diseases.

BACKGROUND: The BNT162b2 mRNA COVID-19 vaccine has been administered to children and adolescents with cancer and hematologic diseases since they are at high risk of manifesting severe symptoms if they have COVID-19 infection but the adequate immune response after vaccination in these immunocompromised patients are questionable. OBJECTIVE: To evaluate the immune response of children and adolescents with cancer and hematologic diseases after receiving 2 doses of the BNT162b2 mRNA COVID-19 vaccine. METHODS: This is a prospective cohort study of patients with cancer and hematologic disease, who aged 12- 18 years old and received 2 doses the BNT162b2 vaccines at 4 weeks apart were enrolled. Immunogenicity was determined by measuring serum anti-SARS-CoV-2 immunoglobulin antibodies directed against the receptor binding domain (RBD) of S1 domain of the spike protein (Anti S-RBD), surrogated viral neutralization test (sVNT) of SARS-CoV-2 and Delta strain. Blood samples were collected and analyzed at 4 and 12 weeks after vaccination. The seroprotective rate was defined as sVNT ≥ 68%. RESULTS: From Oct 2021 to Jan 2022, 43 children were enrolled, 21 were on-therapy and 22 were off-therapy. 25 were hematologic malignancy, 15 solid tumor and 3 hematologic diseases with immunosuppressive drugs. The GMT (95%CI) of a anti S-RBD IgG level at 4 weeks after vaccination were 56.05 (13.2,238.2) and 3633 (2689,4908) BAU/mL in on-therapy and off-therapy group, respectively, p<0.001. The sVNT (95%CI) of delta strain were 26% (5.85-73.55%) and 97.05% (96.0-97.4%) as the seroprotective level which were 33.3% in on-therapy group and 100% in off-therapy group (p<0.001). 14 children in on-therapy group need an additional dose. CONCLUSION: After complete vaccination, the seroprotective rate and antibody level in pediatric and adolescent patients with cancer and hematologic disease who receive immunosuppressive agents are quite low, compared with patients who had complete treatment. Additional dose of primary series should be offered.

Comparison of the Immune Response After an Extended Primary Series of COVID-19 Vaccination in Kidney Transplant Recipients Receiving Standard Versus Mycophenolic Acid-sparing Immunosuppressive Regimen.

Two doses of coronavirus disease 2019 vaccination provide suboptimal immune response in transplant patients. Mycophenolic acid (MPA) is one of the most important factors that blunts the immune response. We studied the immune response to the extended primary series of 2 doses of AZD1222 and a single dose of BNT162b2 in kidney transplant patients who were on the standard immunosuppressive regimen compared to those on the MPA-sparing regimen. Methods: The kidney transplant recipients who were enrolled into the study were divided into 2 groups based on their immunosuppressive regimen. Those on the standard immunosuppressive regimen received tacrolimus (TAC), MPA, and prednisolone (standard group). The patients in the MPA-sparing group received mammalian target of rapamycin inhibitors (mTORi) with low dose TAC plus prednisolone (MPA-sparing group). The vaccination consisted of 2 doses of AZD1222 and a single dose of BNT162b2. Results: A total of 115 patients completed the study. There were 76 (66.08%) patients in the standard group and 39 (33.91%) patients in the MPA-sparing group. The overall median anti-severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) S antibody level at 4 wk after vaccine completion was 676.64 (interquartile range = 6.02-3644.03) BAU/mL with an 80% seroconversion rate. The MPA-sparing group achieved higher anti-SARS-CoV-2 S antibody level compared to the standard group (3060.69 and 113.91 BAU/mL, P < 0.001). The seroconversion rate of MPA-sparing and standard groups were 97.4% and 71.1%, respectively (P < 0.001). The anti-HLA antibodies did not significantly increase after vaccination. Conclusions: The extended primary series of 2 doses of AZD1222 and a single dose of BNT162b2 provided significant humoral immune response. The MPA-sparing regimen with mTORi and low dose TAC had a higher ant-SARS-CoV-2 S antibody level and seroconversion rate compared to the participants in the standard regimen.

Benefits of Switching Mycophenolic Acid to Sirolimus on Serological Response after a SARS-CoV-2 Booster Dose among Kidney Transplant Recipients: A Pilot Study.

Kidney transplant recipients (KTRs) have a suboptimal immune response to COVID-19 vaccination due to the effects of immunosuppression, mostly mycophenolic acid (MPA). This study investigated the benefits of switching from the standard immunosuppressive regimen (tacrolimus (TAC), MPA, and prednisolone) to a regimen of mammalian target of rapamycin inhibitor (mTORi), TAC and prednisolone two weeks pre- and two weeks post-BNT162b2 booster vaccination. A single-center, opened-label pilot study was conducted in KTRs, who received two doses of ChAdOx-1 and a single dose of BNT162b2. The participants were randomly assigned to continue the standard regimen (control group, n = 14) or switched to a sirolimus (an mTORi), TAC, and prednisolone (switching group, n = 14) regimen two weeks before and two weeks after receiving a booster dose of BNT162b2. The anti-SARS-CoV-2 S antibody level after vaccination in the switching group was significantly greater than the control group (4051.0 [IQR 3142.0-6466.0] BAU/mL vs. 2081.0 [IQR 1077.0-3960.0] BAU/mL, respectively; p = 0.01). One participant who was initially seronegative in the control group remained seronegative after the booster dose. These findings suggest humoral immune response benefits of switching the standard immunosuppressive regimen to the regimen of mTORi, TAC, and prednisolone in KTRs during vaccination.

Immunogenicity of a Fractional Dose of mRNA BNT162b2 COVID-19 Vaccine for Primary Series and Booster Vaccination among Healthy Adolescents.

Primary series vaccination with BNT162b2 followed by a booster 5 months later has been recommended for healthy adolescents. We aimed to describe the immunogenicity in a fractional dose of BNT162b2. Adolescents aged 12-18 years were randomized into six arms for primary series administration: 3wPZ30/30 (reference group), 3wPZ30/20, 3wPZ20/20, 6wPZ30/30, 6wPZ30/20, and 6wPZ20/20 µg. A booster was given at 5 months after the second dose using either 10 or 15 µg of BNT162b2. Immunogenicity following vaccination was determined by IgG against receptor-binding domain (anti-S-RBD IgG; BAU/mL), surrogate virus neutralization test (sVNT; %inhibition) and pseudovirus neutralization (pVNT;ID50) against Omicron. Non-inferiority criteria were defined as a lower boundary of the geometric mean ratio (GMR) being greater than 0.67. From September to October 2021, 118 adolescents with a median age (IQR) of 14.9 years (13.9-16.7) were enrolled. Fourteen days after the primary series, the geometric means (GMs) of anti-S-RBD IgG (BAU/mL) were 3090 (95% CI 2761-3460) in 3wPZ30/30. The GMRs of anti-S-RBD were: 0.80 (95% CI 0.67-0.97) in 3wPZ30/20; 1.00 (95% CI 0.83-1.20) in 3wPZ20/20; 1.37 (95% CI 1.13-1.65) in 6wPZ30/30; 1.24 (95% CI 1.02-1.50) in 6wPZ30/20; and 1.36 (1.13-1.64) in 6wPZ20/20. After a booster dose with 15 µg (n = 24) of BNT162b2, sVNT and pVNT against Omicron variant were 91.6 (95% CI 88.4-94.9) and 331 (95% CI 221-495), respectively. In the group that received 10 µg of BNT162b2 (n = 25), sVNT was 85.6 (95% CI 80.0-91.6) and pVNT was 397 (95% CI 267-590). Healthy adolescents had good immune responses to the fractional dose regimen of BNT162b2 and this may be considered as an alternative option.

TIGIT Monoallelic Nonsense Variant in Patient with Severe COVID-19 Infection, Thailand.

A heterozygous nonsense variant in the TIGIT gene was identified in a patient in Thailand who had severe COVID-19, resulting in lower TIGIT expression in T cells. The patient's T cells produced higher levels of cytokines upon stimulation. This mutation causes less-controlled immune responses, which might contribute to COVID-19 severity.

SARS-CoV-2 Omicron variant emerged under immune selection.

The SARS-CoV-2 Omicron variant (B.1.1.529 lineage) escapes antibodies that neutralize the ancestral virus. We tested human serum panels from participants with differing infection and vaccination status using a multiplex surrogate virus neutralization assay targeting 20 sarbecoviruses. We found that bat and pangolin sarbecoviruses showed significantly less neutralization escape than the Omicron variant. We propose that SARS-CoV-2 variants have emerged under immune selection pressure and are evolving differently from animal sarbecoviruses.

Cold agglutinin disease following SARS-CoV-2 and Mycoplasma pneumoniae co-infections.

SARS-CoV-2 and other respiratory co-infections may occur. As Mycoplasma pneumoniae and various viruses can cause cold agglutinin disease (CAD), the presence of CAD in COVID-19 patients should indicate the need of investigations for those pathogens.

A Randomized Clinical Trial of a Fractional Low Dose of BNT162b2 Booster in Adults Following AZD1222.

In the era of globally predominant omicron strains, a COVID-19 booster vaccine is needed. Our study aimed to evaluate the immunogenicity of a half-dose BNT162b2 booster after AZD1222 in healthy adults. A randomized trial of volunteers aged 18-69 years who received two-dose AZD1222 was conducted. The participants were randomized to receive the BNT162b2 vaccine intramuscularly-half (15 µg) vs. standard dose (30 µg). The immunogenicity was evaluated by a surrogate virus neutralization test (sVNT) against omicron variants and anti-spike-receptor-binding-domain IgG (anti-S-RBD IgG). From November-December 2021, 100 adults with a median age of 59.3 years (IQR 33.4-65.5) were enrolled. A booster dose was given at median of 98 days (IQR 92-128) after AZD1222. At day 14, the geometric means (GMs) of anti-S-RBD IgG in half- vs. standard-dose group were 2329.8 vs. 2574.7 BAU/mL, with a geometric mean ratio (GMR) of 0.90 (0.77-1.06). The GMs of sVNT against the omicron variant in the half- and standard-dose groups were 74.4% inhibition (95% CI 68.8-80.5) and 67.3% inhibition (57.9-78.1), respectively, with GMR of 0.95 (0.69-1.30). At day 90, the sVNT indicated 22.3% inhibition (95% CI 14.9-33.4) and 20.4% inhibition (13.1-32.0), respectively, with GMR of 1.09 (0.60-1.98). The fractional low-dose BNT162b2 mRNA booster vaccine provided non-inferior immunogenicity responses. During a shortage of vaccine supply, a fractional low dose should be considered for a booster vaccination program.

Dynamics of Neutralizing Antibody and T-Cell Responses to SARS-CoV-2 and Variants of Concern after Primary Immunization with CoronaVac and Booster with BNT162b2 or ChAdOx1 in Health Care Workers.

Inactivated SARS-CoV-2 vaccine (CoronaVac) is commonly used in national immunization programs. However, the immune response significantly declines within a few months. Our study assessed the immune response against SARS-CoV-2 after receiving booster shots of BNT162b2 or ChAdOx1 among health care workers who previously received CoronaVac as their primary immunization. Fifty-six participants who received ChAdOx1 and forty-two participants who received BNT162b2 were enrolled into this study, which evaluated immune responses, including anti-SARS-CoV-2 spike total antibodies (Elecsys®), surrogated viral neutralization test (sVNT) to ancestral strain (cPass™; GenScript), five variants of concern (Alpha, Beta, Gamma, Delta, and Omicron) (Luminex; multiplex sVNT) and the ELISpot with spike (S1 and S2) peptide pool against the ancestral SARS-CoV-2 strain. The samples were analyzed at baseline, 4, and 12 weeks after primary immunization, as well as 4 and 12 weeks after receiving the booster. This study showed a significant increase in anti-SARS-CoV-2 spike total antibodies, sVNT, and T-cell immune response after the booster, including against the Omicron variant. Immune responses rapidly decreased in the booster group at 12 weeks after booster but were still higher than post-primary vaccination. A fourth dose or a second booster should be recommended, particularly in health care workers.

Host genetic factors of COVID-19 susceptibility and disease severity in a Thai population.

Host genetic factors have been shown to play a role in SARs-CoV-2 infection in diverse populations. However, the genetic landscape differs among various ethnicities; therefore, we explored the host genetic factors associated with COVID-19 disease susceptivity and disease severity in a Thai population. We recruited and genotyped 212 unrelated COVID-19 Thai patients and 36 controls using AxiomTM Human Genotyping SARs-COV-2 array, including 847,384 single nucleotide polymorphisms related to SARs-COV-2 pathogenesis, immune response, and related comorbidity No SNPs passed the genome-wide significance threshold of p value <1 × 10-8. However, with a threshold of p value <1 × 10-5, a locus on chromosome 5q32 was found to have a suggestive association with COVID-19 disease susceptibility (p value 6.9 × 10-6; Q-Q plot λ = 0.805, odds ratio 0.02). Notably, IL17B is a gene located in this linkage disequilibrium block and is previously shown to play a part in inflammation and pneumonia. Additionally, a suggestive locus on chromosome 12q22, harboring EEA1 and LOC643339, was associated with COVID-19 disease severity (p value 1.3 × 10-6 - 4.4 × 10-6, Q-Q plot λ = 0.997, odds ratio 0.28-0.31). EEA1 is involved in viral entry into cells, while LOC643339 is a long non-coding RNA. In summary, our study suggested loci on chromosomes 5q32 and 12q22 to be linked to COVID-19 disease susceptibility and disease severity, respectively. The small sample size of this study may lessen the likelihood that the association found is real, but it could still be true. Further study with a larger cohort is required to confirm these findings.

COVID-19 vaccine acceptance, hesitancy, and determinants among physicians in a university-based teaching hospital in Thailand.

BACKGROUND: The COVID-19 vaccines provide renewed hope in the fight against the recent pandemic. To ensure widespread vaccination, it is crucial to analyze vaccine willingness and its determinants among physicians, key health care influencers. This study aimed to assess acceptance rate and identify factors associated with vaccine hesitancy among Thai physicians. METHODS: A cross-sectional online-based questionnaire was distributed to all physicians at King Chulalongkorn Memorial Hospital during March 31, 2021 to April 30, 2021 in order to assess their attitudes toward receiving the COVID-19 vaccine. Reasons for vaccine acceptance and refusal as well as predictors of vaccine hesitancy were analyzed by bivariate and multivariable analysis. RESULTS: A total of 705 complete responses were received with 95.6% (n = 675) of physicians expressing willingness to receive a COVID-19 vaccine. Only one of the 31 physicians (4.4%) who expressed a hesitancy or unwillingness to be vaccinated was a faculty member; the others were physicians-in-training. Approximately one-fifths of physicians surveyed were also not willing to recommend the vaccine to their family members (21.4%, n = 151) or patients (18.7%, n = 132). Using multivariable logistic regression, vaccine hesitancy was independently associated with preference for particular vaccines over the government allocated option, especially for mRNA vaccine (aOR 8.86; 95% CI 1.1-71.54; p = 0.041). Vaccine literacy showed an inverse relationship (aOR 0.34; 95% CI 0.13-0.9; p = 0.029) with vaccine hesitancy. Uncertainty of the vaccine efficacy (83.9%) and fear of adverse events (48.4%) were major concerns contributing to vaccine hesitancy. CONCLUSION: This study revealed a high rate of physician willingness to take the COVID-19 vaccine especially among staffs; however, a significant proportion would not currently suggest vaccination to their families or patients. Restrictions on vaccine choice and vaccine illiteracy, together with concerns over adverse effects and uncertainty of efficacy, were associated with negative attitudes toward vaccination. To raise acceptance of the vaccination program, efforts should be made to balance individual preference for vaccine type in addition to increasing the availability of accurate data on safety and efficacy for each vaccine.

Clinical outcomes of pediatric COVID-19 in a tertiary care center in Bangkok, Thailand.

Objective: To describe the clinical characteristics and outcomes of pediatric COVID-19 in Thailand, where favipiravir is the mainstay of antiviral treatment. Methods: We conducted a hospital based observational cohort study of COVID-19 among children. The study included children (age <15 years) with confirmed positive reverse transcriptase-polymerase chain reaction for SARS-CoV-2 from nasopharyngeal swab. Results: From April to July 2021, 416 cases with a median age of 7.1 (interquartile range 2.7-11.6) years were included in the study. The spectrum of disease included 82 (20%) asymptomatic, 232 (56%) mild and 102 (24%) with pneumonia. Abnormal chest x-ray findings included ground-glass opacities (46%), focal infiltrations (27%), perihilar opacities (19%), reticular infiltrations (15%) and other non-specific findings (4%). Only 12 children (3%) required oxygen support. Favipiravir was prescribed to 129 children (31%); 102 patients with pneumonia and 27 patients at risk for disease progression. Pneumonia was more common in age <3 years compared with those aged 3-<12 years (adjusted odds ratio (aOR) 0.30, 95% CI 0.17-0.52), 12-15 years (aOR 0.40, 95% CI 0.21-0.77) and in patients with comorbidities (aOR 2.36, 95% CI 1.09-5.12). Conclusions: One-fourth of pediatric COVID-19 patients had pneumonia, but few required oxygen support. Off-label use of Favipiravir in pediatric COVID-19 patients in a recent outbreak in Bangkok is reported.

Short-term immune response after inactivated SARS-CoV-2 (CoronaVac®, Sinovac) and ChAdOx1 nCoV-19 (Vaxzevria®, Oxford-AstraZeneca) vaccinations in health care workers.

BACKGROUND: Inactivated SARS-CoV-2 (CoronaVac®, Sinovac, or SV) and ChAdOx1 nCoV-19 (Vaxzevria®, Oxford-Astra Zeneca, or AZ) vaccines have been administered to the health care workers (HCWs). OBJECTIVE: To determine the short-term immune response after the SV and AZ vaccinations in HCWs. METHODS: In this prospective cohort study, HCWs who completed a 2-dose regimen of the SV or AZ were included. Immune response was evaluated by surrogate viral neutralization test (sVNT) and anti-SARS-CoV-2 total antibody. Blood samples were analyzed at 4 and 12 weeks after the complete vaccination. The primary outcome was the seroconversion rate at 4-weeks after complete immunization. RESULTS: Overall, 185 HCWs with a median (IQR) age of 40.5 (30.3-55.8) years (94 HCWs in the SV group and 91 in the AZ group) were included. At 4 weeks after completing the SV vaccination, 60.6% (95%CI: 50.0-70.6%) had seroconversion evaluated by sVNT (≥ 68% inhibition), comparable to the patients recovered from mild COVID-19 infection (69.0%), with a rapid reduction to 12.2% (95%CI: 6.3-20.8) at 12 weeks. In contrast, 85.7% (95%CI: 76.8-92.2%) HCWs who completed two doses of the AZ for 4 weeks had seroconversion, comparable to the COVID-19 pneumonia patients (92.5%), with a reduction to 39.2% (95%CI: 28.4-50.9%) at 12 weeks. When using the anti-SARS-CoV-2 total antibody level (≥ 132 U/ml) criteria, only 71.3% HCWs in the SV group had seroconversion, compared to 100% in the AZ group at 4 weeks. CONCLUSIONS: A rapid decline of short-term immune response in the HCWs after the SV vaccination indicates the need for a vaccine booster, particularly during the ongoing spreading of the SARS-CoV-2 variants of concern.

Prevalence of thrombocytopenia, anti-platelet factor 4 antibodies and D-dimer elevation in Thai people After ChAdOx1 nCoV-19 vaccination.

BACKGROUND: Vaccine-induced immune thrombotic thrombocytopenia (VITT) is a rare distinctive syndrome characterized by unusual site thrombosis accompanied by thrombocytopenia after ChAdOx1 nCoV-19 vaccination. Platelet-activating anti-platelet factor 4-dependent antibodies (anti-PF4 Abs) were detected in most cases of VITT. To date, data from Asian countries are lacking. OBJECTIVES: To determine the prevalence of thrombocytopenia, anti-PF4 Abs, and D-dimer elevation in Thai people administered the ChAdOx1 vaccine. PATIENTS/METHODS: A total of 521 vaccinated and 146 nonvaccinated subjects were enrolled. Blood samples were collected to determine platelet counts, anti-PF4 Abs using ELISA and D-dimer levels 5 to 30 days after the first vaccination. RESULTS: None of the participants developed thrombocytopenia or had significantly decreased platelet counts from baseline after ChAdOx1 vaccination. The frequencies of anti-PF4 Abs between vaccinated (16/521; 3.1%; 95% confidence interval [CI], 1.8-4.9) and nonvaccinated Thai people (6/146; 4.1%; 95% CI, 1.5-8.7) were similar. None of the detectable anti-PF4 Abs activated platelets in vitro. The average D-dimer levels between vaccinated and control groups were similar (282.2 ± 286.3 vs 267.8 ± 219.3 ng/mL; P = 0.58). Four vaccinated and one nonvaccinated participants had markedly elevated D-dimer levels >2000 ng/mL without detectable anti-PF4 Abs. Imaging studies of these asymptomatic subjects revealed incidental pulmonary embolism in a vaccinated elderly woman. CONCLUSIONS: This study demonstrated a low prevalence of thrombocytopenia and pathogenic anti-PF4 Abs after ChAdOx1 vaccination. D-dimer testing revealed no significant coagulation activation. Routine tests for platelet counts, anti-PF4 Abs, and D-dimer levels are not recommended for VITT screening without clinical suspicion.

Federated learning for predicting clinical outcomes in patients with COVID-19.

Federated learning (FL) is a method used for training artificial intelligence models with data from multiple sources while maintaining data anonymity, thus removing many barriers to data sharing. Here we used data from 20 institutes across the globe to train a FL model, called EXAM (electronic medical record (EMR) chest X-ray AI model), that predicts the future oxygen requirements of symptomatic patients with COVID-19 using inputs of vital signs, laboratory data and chest X-rays. EXAM achieved an average area under the curve (AUC) >0.92 for predicting outcomes at 24 and 72 h from the time of initial presentation to the emergency room, and it provided 16% improvement in average AUC measured across all participating sites and an average increase in generalizability of 38% when compared with models trained at a single site using that site's data. For prediction of mechanical ventilation treatment or death at 24 h at the largest independent test site, EXAM achieved a sensitivity of 0.950 and specificity of 0.882. In this study, FL facilitated rapid data science collaboration without data exchange and generated a model that generalized across heterogeneous, unharmonized datasets for prediction of clinical outcomes in patients with COVID-19, setting the stage for the broader use of FL in healthcare.

Multiple clades of SARS-CoV-2 were introduced to Thailand during the first quarter of 2020.

In early January 2020, Thailand became the first country where a coronavirus disease 2019 (COVID-19) patient was identified outside China. In this study, 23 whole genomes of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) from patients who were hospitalized from January to March 2020 were analyzed, along with their travel histories. Six lineages were identified including A, A.6, B, B.1, B.1.8, and B.58, among which lineage A.6 was dominant. Seven patients were from China who traveled to Thailand in January and early February. Five of them were infected with the B lineage virus, and the other two cases were infected with different lineages including A and A.6. These findings present clear evidence of the early introduction of diverse SARS-CoV-2 clades in Thailand.